Stance Detection on Cancer related Tweets
The list of todos are:
-
Increase the annotation of the dataset for the 3 major categories : X_causes_cancer, X_cures_cancer, X_prevents_cancer (at least 1000 for each category).
-
Develop some initial classification model for identifying relevant tweets for each of these 3 classes (a binary classifier that says whether tweets have the particular information in them or not.)
-
Given the tweets which have medical relevance, extract the information from the tweets about the object that causes, cures, prevents cancer.
-
Finally identify the stance of the object in each of this regard.